Molecular dynamics simulations and in vitro studies of hybrid decellularized leaf-peptide-polypyrrole composites for potential tissue engineering applications

Abstract Tissue engineering (TE) aims to repair and regenerate damaged tissue by an assimilation of optimal combination of cells specific to the tissue with an appropriate biomaterial. In this work, a new biomaterial for potential cardiac TE applications was developed by utilizing a combination of in silico studies and in vitro experiments. Molecular dynamics (MD) simulations for the formation of the novel composite prepared from the decellularized leaf components cellulose and pectin along with the VEGF derived peptide (NYLTHRQ) and polypyrrole (PPy) was carried out to assess self-assembly, mechanical properties, and interactions with integrin and NPR-C receptors which are commonly found in cells of cardiac tissue. Results of molecular dynamics simulations indicated the successful formation of stable assemblies. MD simulations also revealed that the scaffold successfully interacted with integrin and NPR-C receptors. As a proof of concept, beet leaves were decellularized (DC) and cross-linked with NYLTHRQ and PPy using layer-by-layer assembly. Decellularization (DC) was confirmed by DNA and protein quantification. Incorporation of the NYLTHRQ peptide and polypyrrole was confirmed by FTIR spectroscopy and SEM imaging. The DC-NYLTHRQ-PPy scaffold was seeded with co-cultured cardiomyocytes and vascular smooth muscle cells. The scaffold promoted cell proliferation and adhesion. Actin and Troponin T immunofluorescence staining showed the presence of these critical cardiomyocyte markers. Thus, for the first time we have developed a decellularized leaf-peptide-PPy composite scaffold by a combination of in silico studies and laboratory analyses that may have potential applications in cardiac TE. Communicated by Ramaswamy H. Sarma


Introduction
Cardiovascular diseases are one the most common causes of mortality in the world (Laslett et al., 2012). Among these, heart failure due to myocardial infarction (MI) is one of the most serious ailments (Thygesen et al., 2007). The process of tissue repair following MI consists of inflammation, phagocytosis, and scar formation or fibrosis, however cardiac tissue often fails to regain normal function in the affected area due to limited regenerative capacity (Isomi et al., 2019;Prabhu & Frangogiannis, 2016;Sutton & Sharpe, 2000). Long-term effects of MI include congestive heart failure, ruptured vasculature, and cardiogenic shock, resulting in a need for a heart transplantation (Cinq-Mars et al., 2015). However, this form of treatment is often plagued by shortcomings such as limited organ availability, donor site injury, and in some cases, morbidity, deformity, scarring, inflammation, and blood loss (Giannoudis et al., 2005). To circumvent these difficulties, tissue engineering has emerged as a potential alternative form of treatment. Tissue engineering (TE) aims to repair and regenerate damaged tissue by ideal combination of the patient's own cells with an appropriate biomaterial (Weigel et al., 2006). In general, TE is based on the utilization of "scaffolds", or biomaterials that provide the structural framework for cell adhesion, motility and proliferation, and subsequent tissue formation (Lu et al., 2011;Weber et al., 2011). The ideal biomaterial should be able to mimic the components of native extracellular matrix (ECM), and provide structural integrity to a tissue, regulate key processes such as cell-cell communication, cellular adhesion, motility, and angiogenesis (Nikolova & Chavali, 2019).
In the realm of tissue engineering, molecular dynamics (MD) simulations have garnered interest due to their ability to predict the properties of novel biomaterials and assess their stability and their competency. For example, MD simulations have previously been conducted for N 4 -octanoyl-2 0deoxycytidine nucleic acid-based hydrogels in order to assess its self-assembly and gelation capability thus demonstrating its potential as a biocompatible material for drug delivery and tissue engineering (Angelerou, 2018). In another study, the mechanical properties of graphene nanoribbons (GNRs) with Type I collagen was studied using molecular dynamics. The results showed that the shear force of the biocomposite was improved by the presence of GNR ripples during the simulations, thus showing that the polymer could be reinforced by the GNRs (Ebrahimi et al., 2001). These studies are critical for highlighting the applications of molecular dynamics to the field of tissue engineering. Additionally, MD simulations can also provide information regarding the mechanical properties of systems. For example, properties such as binding affinities under simulated pressure, and elastic properties resulting from a simulated stretching force can be calculated (Isralewitz et al., 2013). The ability to examine the mechanical properties of a system is critical for tissue engineering where the designed scaffold will need contractile abilities. For instance, the elastic moduli of scaffolds consisting of carbon nanotubes with either poly (methyl methacrylate) or polyf(m-phenylenevinylene)-co-[(2,5-dioctoxy-p-phenylene) vinylene]g were predicted with the use of molecular dynamics (Han & Elliott, 2007). In general, MD simulations of biohybrids composed of several components is highly pertinent in tissue engineering, since the composite scaffold can be optimized and better designed, thus, making the experimental biocompatibility studies more streamlined and economical.
A suitable scaffold material for cardiac tissue engineering should be cytocompatible, capable of inducing various biological cues, provide appropriate structural support and mechanical properties, and have an architecture that allows for contractility (Akhyari et al., 2008). Recent cardiac TE efforts have focused on synthetic polymers such as polyglycolic acid (PGA), poly (L-lactide-co-glycolide) (PLGA), polycaprolactone, and poly-L-lactic acid (PLA) (Generali et al., 2017;Li et al., 2019;Liu et al., 2015;Radhakrishnan et al., 2021). Electrospun PLGA fibers incorporated with the peptides derived from laminin (YIGSR and RGD) resulted in significantly enhanced adhesion of cardiomyocytes (Yu et al., 2014). The development of cardiac patches through the use of biomaterials has gained interest because of their potential for direct implantation on cardiac tissue and their electrical and mechanical properties (Izadifar et al., 2018). In particular, cardiac patches that boost mouse embryonic stem cells and cardiovascular progenitors have been shown to exhibit intricate 3 D organization and cell differentiation similar to neonatal myocardial cells (Liau et al., 2011).
An increasingly common natural biomaterial gaining popularity for the design and formation of cardiac patches is decellularized tissue. In decellularization, the ECM is isolated from the cellular components of a tissue by exposure to chemical, enzymatic, or physical agents (Crapo et al., 2011). The ECM maintains its native architecture, and inherent structural, biochemical, and biomechanical cues (Bielli et al., 2018;Gilpin & Yang, 2017). For example, when decellularized porcine myocardium was reseeded with cells, the cells maintained cardiomyocyte-like phenotype and possible endothelialization in locations close to vasculature channels (Wang et al., 2010). An intriguing new approach is decellularization of plant materials. Considering their wide availability, cytocompatibility, and cost-efficient supply, decellularized plant materials may be further considered as a possible biomaterial in tissue engineering applications (Gershlak et al., 2017). Furthermore, plant-based scaffolds can circumvent immunological reactions (Predeina et al., 2020).
To our knowledge, for the first time, in this work, we conducted molecular dynamics simulation studies of a newly designed scaffold that can mimic the components of decellularized leaves using cellulose and pectin. These components were then attached to VEGF-derived peptide sequence NYLTHRQ to increase vascularity. VEGF (vascular endothelial growth factor) has been extensively investigated in TE applications to enhance myocardial repair and function (Des Rieux et al., 2011). This sequence in particular, is well-known for its angiogenic properties and was found to directly support endothelial cell adhesion through its interaction with alpha-5-beta-1 integrin (Soro et al., 2008). To impart electrical properties to the matrix, the conductive polymer polypyrrole (PPy) was then incorporated. In previous work, PPy has been shown to be biocompatible and has been utilized to promote electrical conduction in neural and cardiac tissue scaffolds (Huang et al., 2014;Runge et al., 2010;Talebi et al., 2020). For example, hybrid polypyrrole scaffolds with alginate, chitosan or collagen have been developed due to their desirable properties such as electrical conductivity, porosity, and biodegradability for preparation of neural scaffolds (Manzari-Tavakoli et al., 2020;Yow et al., 2011). In a recent study it was shown that polypyrrole encapsulated silk fibroin nanofibers were able to closely mimic the mechanical properties of myocardium and demonstrated electrical conductivity and supported cardiomyocyte contraction (Liang et al., 2021). Thus, we incorporated polypyrrole to the cellulosepectin-NYLTHRQ assemblies.
Molecular dynamics (MD) simulations were used to examine the formation of the scaffold assembly after each layer was added. These assemblies appeared to have a transient nature, frequently reordering over time, until all molecules attached together as single, flexible fibrous assembly. The RMSD values indicated stability of the scaffold and the density was found to increase, indicating closer interactions among the individual components. In addition, the scaffold was shown to interact with and assemble around the natriuretic peptide receptor-C and alpha-5-beta-1 integrin receptor which are commonly expressed in cardiomyocytes and are essential for attachment of scaffolds to the extracellular matrix of the cells. Because MD simulations have been shown to predict the behavior of biomaterials in vitro (Migliavacca et al., 2005), as a proof of concept, we utilized decellularized beet leaves (which are rich in cellulose and pectin) as a template for preparation of the scaffold. To the decellularized leaf components, we next incorporated the peptide NYLTHRQ, followed by PPy and examined the impact of the scaffold on co-cultured cardiomyocytes and vascular smooth muscle cells. Our results indicate that the scaffolds are biocompatible, and promote cell proliferation and adhesion of co-cultured cells. Furthermore, the presence of biomarkers for myofibrillar proteins expressed by cardiomyocytes and smooth muscle cells that include Cardiac Troponin T and alpha actinin was also revealed. Thus, we modeled and developed a functional, multi-layered scaffold with desirable properties for potential applications in cardiac TE.

Materials
Fresh beet with stem and leaves intact (beta vulgaris) was purchased from a local market. Sodium Dodecyl Sulfate (SDS), Phosphate buffer Saline (PBS, pH 7.4), 4% formaldehyde, Triton-X-100, Hanks Balanced Salt solution (HBSS) and sodium hypochlorite bleach were purchased from Fisher Scientific. NYLTHRQ was custom ordered from Genscript. Cardiomyocyte isolation kit was purchased from Pierce Biotechnology (Rockford, IL, USA). Vascular smooth muscle cell growth kit (ATCC-PCS-100-042) and Vascular Cell Basal Medium were purchased from ATCC (Manassas, VA, USA). Rat primary artery smooth muscle cells and neonatal hearts were purchased from Cell Biologics. Cardiomyocyte growth media and growth factors were purchased from Science Cell. Bisbenzimide DNA quantification kit, Polypyrrole (undoped, $20 wt. % loading, composite with carbon black), acetonitrile, Bradford reagent, BSA were purchased from Sigma Aldrich. The MTT Assay kit was purchased from Cayman Chemical Company (Ann Arbor, MI, USA). Primary and secondary antibodies for cardiac Troponin T and actinin assays were purchased from Santa Cruz Biotechnology.

Structure design
Chemical structures were designed using ChemDraw and ChemDraw 3 D (19.1). Individual molecules of cellulose, pectin, NYLTHRQ, and PPy were prepared first in ChemDraw, then transferred to ChemDraw 3 D (19.1) in order to minimize the energy of the structures. This was done utilizing the "MM2" minimize energy calculation on ChemDraw 3 D (19.1). The structures were then exported as .pdb files and imported into Maestro 12.5 for visualization and molecular dynamics simulations.

Molecular dynamics
To assess the formation of the scaffold and its assembly as well as to examine its mechanical and elastic properties, molecular dynamics simulations were conducted. The simulations were performed using DESMOND through the software Maestro 12.5 [Maestro, Schrodinger, LLC, New York, NY. (2020). Desmond Molecular 1683 Dynamics System, D. E. Shaw Research, New York, NY, 2020. Maestro-1684 Desmond Interoperability Tools, Schroedinger, New York, NY]. The systems for molecular dynamics simulations were built using the disordered system builder in Maestro 12.5. The systems contained an even distribution of each molecule. To construct the scaffold, we initially designed a system containing 30 molecules of cellulose and 30 molecules of pectin, which represented the components of decellularized leaves. To that we added 30 molecules of the peptide NYLTHRQ. The final simulation consisted of an additional layer of 30 molecules of PPy with the decellularized leaf components and peptide. The simulations were run for 200 ns while utilizing a compressive relaxation protocol and a time-step of two fs.
Following the simulation, the resulting scaffolds were subjected to stress-strain property calculations using Maestro 12.5. The calculation utilized the volume-conserving uniaxial method which provided strain along the axes with a step size of 0.015 e over 40 steps and ran for 100 ps with a time step integration of 2.0 fs. The strain rate was maintained at 1.5 Â 10 8 s À1 . With this method, the stress was calculated at each increment by determining changes in the pressure tensor between the initial and deformed states. The temperature was kept at 300 K and the stress averaged over the final 20% of the trajectory during which time the system is predicted to be stable. A stress-strain curve was generated to determine the Young's modulus for each system.
We also utilized molecular dynamics to examine the interactions of the scaffolds with receptors expressed in cardiomyocytes. The receptors were obtained from the RCSB database and prepared using the protein preparation wizard on Maestro 12.5. The receptors utilized were the, alpha-5beta-1 integrin (RCSB ID: 5HJ2) (Brown et al., 2018) and natriuretic peptide receptor-C (NPR) (RCSB ID: 1YK0) (He et al., 2006). Any ligands already bound to the crystallized structures were removed prior to the simulations in order to assess the interactions between the scaffold and receptor without ligand interference. These simulations were performed in an SPC model with a 0.15 M NaCl solution to mimic physiological salt concentrations.
2.2.3. Preparation of scaffold 2.2.3.1 Decellularization of beet leaves. The leaves were first washed with deionized (DI) water, and the stems were trimmed off. Decellularization techniques were adapted from previous methods (Gershlak et al., 2017). Briefly, the leaves were soaked for ten minutes in hexane to remove the cuticles and then soaked in phosphate buffer saline (PBS) solution for 10 minutes. Leaves were then immersed in with 10% SDS solution for 7 days, after which they were removed from SDS, washed with DI water, and placed in a solution of 10% Triton-X-100 for 30 minutes. They were when washed thoroughly in DI water and allowed to soak in 6% sodium hypochlorite bleach for an hour. The leaves were again thoroughly washed with DI water, and stored in DI water at 4 C.

Sample preparation for DNA/protein quantitation.
Decellularized and native leaves were placed in cryogenic vials and stored in a liquid nitrogen bath for 15 minutes. Leaves were then removed from vials and immediately put into a mortar-pestle to form a paste. The paste was then diluted with DI water and sonicated for 10 minutes. The samples were subsequently centrifuged at 15000 rpm for 5 minutes. Then 800 mL of the supernatant was pipetted into an Eppendorf tube and centrifuged for another 3 minutes. The resulting supernatant was used for DNA and protein quantification. DNA quantification was carried out using bis benzimide H-33258 fluorescence-based assay in 384 well plates and analyzing using a Fluodia T70 from PTI fluorescence plate reader. Protein quantification was performed using Bradford assay in 96-well plates using a Gen5 microplate reader (BioTek Instruments). All analyses were carried out in triplicates. The concentrations of standard solutions of protein were determined using Bradford assay using BSA standard.

Cross-linking with peptide and polypyrrole.
NYLTHRQ (1 mg/mL) in pH 7 was prepared and stored at 4 C. Decellularized leaves were cut into 2 cm x 2 cm keeping the midsection of the leaf and washed with PBS. Then 1% glutaraldehyde was added dropwise throughout the surface of the leaf in order to allow for cross-linking, and allowed to sit for 12 hours at 4 C. Then the leaves were washed with DI water and incubated in a solution containing 10 mL of NYLTHRQ (1 mg/mL) with mild shaking at 50 rpm. The peptide was allowed to bind to decellularized beet leaves for 48 hours at 4 C. To coat with polypyrrole in the next step, briefly, the peptide coated leaves were air dried and allowed to incubate in 500 mL of polypyrrole (1 mg/mL) prepared in 50% acetonitrile-water and allowed to bind for 48 hours. The peptide-polypyrrole bound scaffolds were then washed with PBS and stored at 4 C for further analysis. For FTIR analysis, several 2 cm x 2 cm decellularized leaves (after incorporation of peptide and then after incorporation of polypyrrole) were crushed using a mortar and pestle and the paste obtained was lyophilized and stored before further analysis.

Scanning electron microscopy (SEM)
To examine morphologies, decellularized beet leaves (DC), peptide coated decellularized beet leaves (DC-pep), peptide and polypyrrole coated decellularized beet leaves (DC-pep-PPy) were imaged using SEM. For fixing samples, they were trimmed to 5 mm x 5 mm pieces and placed in 3 mL of 1.5% glutaraldehyde and sodium cacodylate (0.7 M) for 2 hours. Subsequently, the glutaraldehyde/sodium cacodylate solution was pipetted out and leaves were washed with serial dilutions of ethanol (starting with 30% concentration and progressing to 50%, 80%, and 95% ethanol). Samples were left in each ethanol concentration for 5 minutes before progressing to the next concentration. The samples were stored in a refrigerator until completely dried. SEM imaging was performed using a Zeiss EVO MA10 model (Oberkochen, Germany).

Fourier transform infrared (FTIR) spectroscopy
To further confirm incorporation of each layer, FTIR spectroscopy was conducted using a Thermo Scientific, Nicolet IS50 FTIR (Thermo Scientific) and analyzed using OMNIC Software. Measurements were taken between 400 and 4000 cm À1 with 10 scans per sample which were then averaged.

Isolation of cardiomyocytes from neonatal rat heart
Cardiomyocytes (CMCs) were isolated using the protocols given in the isolation kit. Briefly, rat hearts were minced into 1-3mm 2 pieces and washed with ice cold HBSS (500 mL) twice to remove blood. Reconstituted cardiomyocyte isolation enzyme 1 (200 mL) (with papain) and cardiomyocyte isolation enzyme 2 (10 mL) (with thermolysin) was added to the minced hearts and incubated at 37 C for 30 minutes. Next, the enzyme solution was removed and the tissues were washed twice with ice cold HBSS (500 mL). Then complete DMEM (500 mL) (containing 10% fetal bovine serum and 1% final concentration of penicillin-streptomycin) was added to the tissue, and vortexed to further break up the tissue. Once in cell suspension, the cells and media were transferred into a 25 mm 2 culture flask and additional media was added to the flask and cells were allowed to grow to confluence at 37 C in a humidified 5% CO 2 atmosphere incubator. Cardiomyocyte growth factors were then added to the flask after 24 hours.

Cell culture
Vascular smooth muscle cells (SMCs) were grown to confluence in 75 cm 2 culture flasks in vascular cell basal medium supplemented with 5 ng/mL rh FGF-basic, 5 mg/mL rh insulin, 50 mg/mL absorbic acid, 10 mM L-glutamine, 5 ng/mL rh EGF, 5% fetal bovine serum, and 10% penicillin/streptomycin and ampicillin mixture at 37 C in a 5% CO 2 humidified atmosphere. The growth of cells was monitored by optical microscopy every 24 hours and cells were split every 48-72 hours. After CMCs and SMCs were grown to confluence individually at a cell density of 2 Â 10 5 cells, co-cultures containing both cell lines were then allowed to grow in a 25 mm 2 culture flask containing 1:1 ratio of CMCs and SMCs in growth media supplemented with growth factors containing 1 Â 10 5 CMCs and 1 Â 10 5 SMCs.

Cell viability (MTT assay)
To examine cell viability in the presence and absence of the DC-pep-PPy scaffolds (3 mm x 3 mm) in 96-well Falcon polystyrene tissue culture plates. We added DC-pep-PPy scaffolds to the wells and plated cells at a density of 2 Â 10 4 cells/ well. The cells were allowed to spread and attach for 24 hours. The growth of the cells and metabolic activity was monitored over a period of 1, 4, and 11 days. To determine cell viability, we performed MTT assay. The absorbance at 570 nm was monitored at each time point using a BioTek Eon microplate reader. Triplicate experiments were run in all cases. The absorbance of media alone was used as the blank and was subtracted from all samples. 2.4.4. Immunofluorescence assays 2.4.4.1. Troponin T assay. DC-pep-PPy scaffold was added to each well of poly-L-lysine coated 24 well falcon tissue culture plates. Co-cultures of CMCs and SMCs were then seeded at a cell density of 2 Â 10 4 cells/well. The DC-pep-PPy scaffolds were cut into the specific dimensions of 3 mm x 3 mm before incubation. The media was changed every 48 hours. After eleven days the media was aspirated off and the cells were washed in PBS. Samples were then fixed in 4% paraformaldehyde overnight and washed with PBS. To permeabilize the cells, samples were washed with 0.5% Triton X for 10 minutes and then washed in PBS and blocked for one hour using 10% bovine serum albumin (BSA). Next the samples were treated with primary cardiac Troponin T antibody and incubated overnight. The unbound primary antibodies were then removed by washing with PBS followed by incubation for three hours with Alexa Fluor 488-conjugated goat anti-mouse IgG polyclonal antibody (Abcam) which was used as the secondary antibody. Samples were then washed with PBS and mounted on to a Amscope Phase Contrast Fluorescence Inverted microscope for imaging.
2.4.4.2. Actinin assay. DC-pep-PPy scaffold was added to each well of poly-L-lysine coated 24 well falcon tissue culture plates. Co-cultures of CMCs and SMCs were then added at a density of 2 Â 10 4 cells/well. The DC-pep-PPy scaffolds were cut into the specific dimensions of 3 mm x 3 mm before incubation. The media was changed every 48 hours. After five days and eleven days, the media was aspirated off and the cells were washed in PBS. Samples were then fixed in 4% paraformaldehyde overnight and washed with PBS. To permeabilize the cells, samples were washed with 0.5% Triton X for 10 minutes and then washed in PBS and blocked for one hour using 10% bovine serum albumin (BSA). The samples were subsequently exposed to alpha-actinin-4 antibody and incubated overnight. Next, the unbound antibodies were removed by washing with PBS followed by incubation for three hours with the secondary antibody, m-IgGj BP-CFL 488. Finally, the samples were washed with PBS and mounted on the fluorescence microscope for imaging.

Statistical analysis
The standard deviations were calculated. Statistically significant differences were then determined using student's t-test. A p value of p < 0.05 was considered statistically significant.

Decellularization
To prepare the scaffold, beet leaves were first decellularized. In general, we observed a change from green color to a muted brown, with increasing translucence indicating loss of chlorophyll. After exposure to Triton-X and bleach, the color was completely removed. In order to confirm decellularization, we examined the DNA and protein content of the decellularized leaves. Since DNA is located in the nuclei of cells, it can provide a good measure of cell removal and, decellularization efficacy (Keller, 1993). Results obtained are shown in Figure 1. The decellularized beet leaves contained less DNA (15 mg DNA/mL tissue) than the native leaves (56 mg DNA/mL tissue); indicative of a 71.9% reduction (Figure 9a). Since proteins are cited as the main cause for immune rejection, protein content of the decellularized leaves was also measured (Cassab, 1998). Major plant cell wall protein groups include hydroxyproline-rich glycoproteins, arabinogalactan proteins, glycine-rich proteins, and proline-rich proteins (Baptista et al., 2018;Jha et al., 2011). We observed an 84.0% reduction in protein content of the decellularized beet leaves compared to the native leaves ( Figure 9b). These results indicate that the decellularization was effective.

Scanning electron microscopy (SEM)
To examine the morphology of the beet leaf derived scaffolds we conducted SEM imaging ( Figure 2). As shown in Figure 2a the decellularized beet leaves appeared to retain venation. The structure was found to be a vastly porous structure with a relatively smooth surface. We also conducted optical microscopy (see supplementary information Figure 1), which indicated also a porous structure. Upon incorporation of the peptide (Figure 2b) we observed a change in the morphology. The surface appeared to have a relatively thick coating on the surface of the decellularized leaf. This is most likely due to the incorporation of the peptide due to cross-linking on the surface and due to H-bonding interactions with the sugar moieties of the cellobiose and pectin components. When PPy was incorporated, we observed a further change in morphology (Figure 2c). Overall aggregates of PPy were found embedded throughout the surfaces of cellulose. Similar aggregates were reported when polypyrrole films were deposited on porphyrin derivatives Figure 1. Comparison of (a) DNA content and (b) protein content before and after decellularization of beet leaves. (Mazurek et al., 2013) and also when cellulose-based electrospun fibers functionalized with polypyrrole (Gorzs as et al., 2011). These results confirm the incorporation of PPy on to cellulose derived from the decellularized beet leaves.

Fourier transform infrared (FTIR) spectroscopy
The incorporation of the protein sequence as well as the polypyrrole into the beet leaf scaffolds was also further confirmed through FTIR spectroscopy. Figure 3 shows the spectra obtained for DC beet, DC beet peptide, and DC beet peptide-PPy scaffolds. Decellularized beet leaf exhibited a broad hydroxyl peak at 3402 cm À1 , while peaks at 2920 cm À1 and 2850 cm À1 due to symmetric and asymmetric aliphatic C-H stretching were observed due to the carbon backbone of cellulose. The peak at 1720 cm À1 is attributed to the ester linkages of the hemicellulose moiety. The peaks at 1470 cm À1 and 1440 cm À1 are attributed to the C-H bending in cellulose or hemicellulose. The peak observed at 1110 cm À1 is due to lignin, however, the peak could be due to stretching vibrations of the C-OH side groups and the C-O-C glycosidic band vibrations of polysaccharides, but is particularly associated with cellulose (Anitha et al., 2019;Sinyayev et al., 2020). Additionally, the peak 1380 cm À1 is due to C-H and -CH 2 bending in either cellulose or hemicellulose.
Incorporation of the peptide resulted in shifts of peaks. The broad -OH peak shifted to 3364 cm À1 . The aliphatic C-H stretch peaks were found to be at 2916 cm À1 and at 2848 cm À1 , while the carbonyl peak was shifted to 1730 cm À1 . The amide I peak was observed at 1630 cm À1 while the amide II peak was observed at 1501 cm À1 after incorporation of the peptide. This indicates interactions between the carbonyl group of the peptide and the ester groups of the cellulose components of the leaves. The peak at 1440 cm À1 was shifted to 1428 cm À1 , and the peak at 1380 cm À1 shifted to 1371 cm À1 . Addition of polypyrrole to the scaffolds caused further changes in the IR spectra. The Amide I peak shifted from 1631 cm À1 to 1650 cm À1 . An additional peak at 1540 cm À1 appeared which can be attributed to the characteristic C ¼ C stretching of the pyrrole rings (Kato et al., 1991). Furthermore, peaks at 1160 cm À1 and 1040 cm À1 were exhibited, possibly due to C-H in plane deformation of the polypyrrole and N-H in plane deformation of the polypyrrole respectively (Eisazadeh, 2007). To further investigate the interactions involved in the self-assembly process, molecular dynamics studies were carried out by mimicking decullarlized plant components, the NYLTHRQ, and PPy.

Self-assembly
Molecular dynamics simulations provide insight into the stability, interactions, and mechanical properties of a designed biomaterial. We assessed these properties using decellularized leaf components, the peptide NYLTHRQ, and PPy. The self-assembly of these components was visualized using the trajectory images generated through Maestro 12.5. The results are shown in Figure 4. The first layer of the scaffold was comprised of 30 molecules each of two major decellularized leaf components, specifically cellulose and pectin (Harris et al., 2021). As shown in Figures 4a-c, there are minor changes in the organization of this layer, with the pectin being most concentrated on the innermost region and cellulose on the outer layer. However, cellulose and pectin remain attached throughout the process and it can be seen that the pectin is integrating in between cellulose fibers. In a previous  study assessing the behavior of cellulose and pectin from beet sugar films, it was found that pectin would integrate into and bind to cellulose microfibrils through the formation of hydrogen bonds and covalent bonds between the ester of pectin and hydroxyl groups of cellulose (Dufresne et al., 1997). The structure of cellulose consists of b (1,4)-D-glucose repeats. In comparison, pectin is made up of a (1,4)-D-galacturonic acid repeats. It is likely that the presence of hydrophilic -OH groups in the polysaccharides as well as -COOH groups of galacturonic acid present in pectin contributed to the formation of stable assemblies, with the chains intricately binding to each other.
To this layer, 30 molecules of the peptide sequence NYLTHRQ were added, exhibiting a trajectory that shows that the peptide attached to the cellulose-pectin assemblies (Figures 4d-f). Changes were observed throughout the assembly process. At 50 ns, the peptide seems to efficiently form layers in the interior portions of the composite. Cellulose is found on the outer layer at 175 ns, while the peptide, and pectin are at the innermost region ( Figure 4f). Furthermore, the entire assembly acquired an oblong shape akin to cross-linked fibrous structures. The peptide NYLTHRQ consists of three polar, uncharged residues, two polar, charged residues, and two hydrophobic residues. Thus, the overall structure is expected to have more hydrophilic properties which would enable it to interact with the hydroxyl and carboxyl moieties on cellulose and pectin, respectively. Further, it has been shown that polar amino acid residues form hydrogen bonds within proteins that have significant function in maintaining and stabilizing protein structure (Worth & Blundell, 2010). Thus, it is possible that the peptide assembling in the interior layer of the scaffold is related to its ability to stabilize polymeric structures. The final layer incorporated to the scaffold was PPy. The trajectory initially shows an even distribution of the scaffold components throughout the 3 D structure in comparison to the previous two simulations (Figures 4g-i). However, at the 50 ns time point PPy appears to be more toward the edges and remains attached toward the edges of the assembly for the simulation process. The structure of PPy consists of a series of imine groups that form polycations. The -NH of the imine group has the potential for hydrogen bonding with the polar components of the cellulose, pectin, and peptide. In addition, PPy is capable of forming stacking interactions with aromatic residues, such as Tyr and with His on the peptide . Such binding interactions enable favorable incorporation of PPy into the scaffold.
Following trajectory analyses, we then assessed the bulk properties of each simulation. The individual graphs of density and volume scaled to each simulation are shown in the Supplementary Material Figures 2-4. The base layer of the scaffold, consisting of cellulose and pectin, exhibited a dramatic reduction in the volume of the system at 30 ns, reaching equilibrium at 60 ns (Supplementary Material Figure S2a). In addition, as a result of the volume reduction, the density of the system was seen to increase, following the trend of the volume changes (Supplementary Material Figure S2b). Upon incorporation of NYLTHRQ, the volume of the assembly is seen to drop immediately during the simulation, reaching equilibrium at 75 ns (Supplementary Material Figure S2a). Further, the density for this simulation increased following the trend of the volume minimization (Supplementary Material Figure S2b). The incorporation of PPy to the scaffold resulted in smaller changes in volume and density. The volume appears to be in an equilibrium state for the duration of the simulation, though it shows a slightly downward trend (Supplementary Material Figure S3a). In a similar manner, the density remains fairly stable with a slight upward trend over the 200 ns (Supplementary Material Figure S3b). A comparison of the bulk properties, including volume, density, and overall stability, of the simulations is shown in Figure 5.
In general, the assemblies show slight decreases in their volumes during the simulations, though they remain fairly steady. These small changes in volume suggest that the components successfully assembled and formed stable complexes. The incorporation of the peptide to the cellulose and pectin components led to a greater volume for the assembly, while the incorporation of PPy led to the smallest volume, indicating that the scaffold was most stable with the layer of PPy added (Figure 5a). Further, the assemblies all show slight increases in density or remain at the same average density during the simulation (Figure 5b). These minimal changes suggest stability.
The self-assembly and incorporation of each layer was determined through analyzing the root mean square deviation (RMSD) during molecular simulation ( Figure 6). The initial layer of cellulose and pectin showed the lowest values for RMSD in the range of 21 A to 31 and reached equilibrium almost immediately (Figure 6a) however slight dips were observed at 60 ns and at 165 ns indicating changes during self-assembly during those time periods, however after 165 ns the system remained stable throughout the simulation. The incorporation of NYLTHRQ was found to slightly increase the RMSD value (overall range from 28 A to 35 A ) (Figure 3b), but similarly reached equilibrium nearly immediately and remained in equilibrium throughout the simulation. No major deviations were  observed during the simulation indicative that incorporation of the peptide sequence resulted in a stable system. Upon incorporation of PPy ( Figure 6c) the RMSD values were found to be in the range of 21.3 A to 33.2 A, which is slightly lower than the second layer. However, more deviations were observed compared at 10 ns, 85 ns and at 141 ns after which no major deviations were observed. This is likely due to additional pi-cation interactions and stacking interactions that come into play upon incorporation of polypyrrole, which may initial disrupt the system, before equilibrating. Since the complete scaffold reached and reached overall, the system can be considered stable. Further, each assembly has small deviations in its RMSD values over the course of the 200 ns simulation and the values remained between 2.1 nm and 3.5 nm overall. RMSD is a measure of atom displacement throughout the simulation, and small changes in RMSD is indicative of a stable structure (Mart ınez, 2015).
Interactions between scaffold components were assessed in addition to the RMSD values. Various interactions were generated during the simulations, including hydrogen bonds, salt bridges, pi-stacking interactions, and pi-cation interactions. These interactions are summarized in Figure 7. The base layer of the scaffold, made up of cellulose and pectin, primarily formed hydrogen bonds during the molecular dynamics simulation. This is mainly due to H-bonding interactions between the -OH groups of b-D-glucose components of cellulose and a-1,4-linked D-galacturonic acid components of pectin.
It is well known that cellulose chains assemble into fibrillar structures due to the formation of inter and intramolecular hydrogen bonding network and hydrophobic effects (Garg et al., 2020). Our results indicated that the range of hydrogen bonds for cellulose and pectin remained within 800-900 hydrogen bonds (Figure 7a). Upon incorporation of NYLTHRQ, a similar number of hydrogen bonds (802-884) was formed as the cellulose and pectin components. Based on these results, it appears that hydrogen bonding interactions played a major role in the self-assembly of these scaffolds. However, the addition of the peptide NYLTHRQ introduced additional interactions to the assemblies and increased the overall number of interactions. The simulation with cellulose, pectin, and NYLTHRQ introduced salt bridges (Figure 7b), pi-stacking, and pi-cation interactions. This is because of the interactions between the carboxyl groups of the poly-galacturonic acid and the amine groups of the peptide (guanidino group of R and N terminal of asparigine), in addition to the presence of tyrosine and histidine which contribute to pi-stacking and pi-cation interactions respectively (Liao et al., 2013). Furthermore, the formation of salt bridges is indicative of interactions between the carboxylic acid component of poly-galacturonic acid and the basic residues of (arginine and histidine) of the peptide (Debiec et al., 2014).
The incorporation of PPy to the scaffold resulted in further changes. An increased number of pi-stacking interactions was  seen. The range of pi-stacking interactions (Figure 7c) was between 122-150 for the duration of the simulation, while the assembly without PPy formed a maximum of 48 pi-stacking interactions. Thus, the binding interactions with PPy are enhanced due to stacking interactions with the tyrosine and histidine residues of the peptide. A similar trend is seen with the pi-cation interactions to a smaller degree. The scaffold with decellularized leaf components and NYLTHRQ formed fewer pication interactions (4-26 interactions) than when PPy was incorporated (16-44 interactions) (Figure 7d). The number of H-bonds was found to be within the range of up to 713. In general, polypyrrole is known to interact with cellulose components through hydrogen bonds between imine groups of pyrrole and hydroxyl groups of cellulose (Al-Dulaimi et al., 2018) as well as contributes to pi-cationic interactions with the peptide components.

Mechanical properties through molecular simulations
Following the scaffold assembly, mechanical properties were analyzed in response to strain applied on the system. The strain-controlled tensile test simulations were conducted by applying uniaxial tension to the scaffold along the x, y, and z axes and assessing the stress applied to the system as a result of the strain. The simulation was repeated for 40 steps total over the course of 100 ps, generating a stress-strain curve. Young's modulus was calculated as the equivalent of the slope of the linear portion of the curve for each simulation. The results (stress-strain curves) are shown in Figure 8. Initially, the components of the decellularized leaves (cellulose and pectin) were analyzed and the Young's modulus was found to be 93 MPa.
Then, the peptide NYLTHRQ was incorporated to examine its effect on the elasticity. The Young's modulus was found to increase to 239.5 MPa due to incorporation of the peptide. As seen, all of the curves follow a similar pattern in which there is initial increases in the amount of stress on the system followed by minimal changes after 0.0512 e of strain is applied. Upon incorporation of PPy, the Young's modulus (YM) was found to further increase slightly and the value obtained was 260.5 MPa. These increases in YM are expected as new layers are added (Park et al., 2008), in particular the increase upon incorporation of the peptide was higher because of a more compact and stable structure formed upon incorporation of the peptide. The authors proposed that the brittle nature of PPy could contribute to this effect, as well as the possibility of PPy interrupting the structure of the polyurethane polymer. It is possible in the present study that the incorporation of PPy and the increase of both pistacking and pi-cation interactions interrupted the hydrogen bonding network of the scaffold formed between cellulose, pectin, and NYLTHRQ, causing a reduction in the cross-linking scaffold.
Enhanced mechanical properties are essential for cardiac tissue due to its necessary contractile function. Thus, the incorporation of PPy provides increased cardiac functionality and electrical properties to the scaffold containing the peptide NYLTHRQ and decellularized leaf components. Previous studies have developed scaffold for cardiac tissue engineering composed of poly (octamethylene maleate (anhydride) 1,2,4-butanetricarboxylate) which maintained elastic properties normally found in cardiac tissues and could be optimized for systems in need of increased elasticity (Davenport Huyer et al., 2016). The Young's modulus of the complete scaffold appears to be in the range observed for carbon nanotubes embedded in poly (glycerol sebacate) and gelatin (Kharaziha et al., 2014) which was found to have imporved contractile properties in the presence of those composites. However, the value obtained in this work is lower than that obtained for TEMPO-oxidized cellulose nanofibers (TOCN), and polypyrrole (PPy) films (Bideau et al., 2016) albeit the Young's modulus has been shown to vary based on the synthetic method used and the thickness of the film. Thus, the value obtained here in is within the range of nanofiber composites utilized for tissue engineering applications.

Molecular dynamics for studying cell-receptor interactions
To assess if the assemblies could interact with cells, the complete scaffold, consisting of cellulose, pectin, the peptide NYLTHRQ, and PPy, was analyzed for its ability to assemble and interact with cell receptors expressed in cardiac tissue. The results are shown in Figure 9. The scaffold was first assembled with the alpha-5-beta-1 integrin receptor, which is a transmembrane glycoprotein that functions to bind cells to the extracellular matrix (Ba c akov a et al. , 2004). Integrins are responsible for transmitting signals between extracellular ligand adhesion sites and the cytoskeleton, which plays a critical role in cell survival, differentiation, and proliferation (Hynes, 2002) and integrin binding materials have been found to have numerous applications in tissue engineering. As shown in the trajectory images, the scaffold begins to move closer to and interacts tightly with the integrin receptor by 50 ns (Figures 9a and b). However, by 175 ns, the appears to anchor to the alpha-subunit of a2b1 integrin. This indicates that the assemblies appear to have an affinity toward integrin and that may aid in developing efficient biomimetc assemblies that can form cell-scaffold matrices. We then explored the interactions NPR-C with the complete scaffold for a 200 ns simulation. NPR-C is associated with signaling pathways in cardiomyocytes that lead to an increased intracellular concentration of cyclic GMP and is involved in pacemaker activities in the cell (Rose & Giles, 2008). This cyclic GMP pathway activated by NPR-C has also been shown to function in the promotion of angiogenesis and vascular remodeling, which are critical for tissue repair (Bubb et al., 2019;. The scaffold with NPR appeared to be relatively stable during the simulation and the scaffold can be seen to assemble around the receptor, evidenced by the images of the trajectory (Figure 6d-f) and remains tightly attached.
After analyzing the trajectory images, the RMSD values were compared. The RMSD values of the scaffold and the integrin receptor complex show greater deviation throughout the simulation when compared to the NPR-C complex ( Figure 10) which is evident from the trajectory images because in the case of integrin the assemblies underwent a change in position at different time points. In the case of integrin, there appears to be dips in RMSD at 18 ns, 70 ns and at 175 ns, likely due to the changes in assembly position with respect to the receptor at those times. However, the RMSD values do appear to reach an equilibrium, indicating stability. The NPR-C RMSD values show minimal deviation and reach equilibrium almost immediately. NPR-C with the scaffold showed overall lower RMSD values compared to the integrin receptor and had smaller deviations indicating higher stability as the assemblies gathered around the receptor. Overall interactions between the scaffold and receptors were assessed, and are summarized in Figure 11.
The scaffold and integrin receptor complex forms hydrogen bonds, salt bridges, pi-pi stacking, and pi-cation interactions. The number of hydrogens bonds ranges from 66-125 throughout the simulation, which is significantly lower compared to the interactions formed among the scaffold alone ( Figure 8a). NPR-C formed a similar number of hydrogen bonds with the scaffold as the integrin receptor, with a range of 65-122 throughout the simulation. The salt bridges (Figure 8b), pi-pi stacking (Figure 8c), and pi-cation interactions ( Figure 8d) are formed only minimally by the integrin receptor complex during the simulation, with ranges of 0-3, 0-1, and 0-1 respectively. NPR-C was also involved in salt bridge formation and pi-cation interactions. The number of salt bridges ranged from 2-34 interactions for NPR-C, which is much larger compared to the integrin receptor complex. The number of pi-cation interactions formed between NPR-C and the scaffold was comparable to the integrin receptor, with a range of 0-3 interactions. The findings of the molecular dynamics simulations suggest that the decellularized leaf components could be improved with the incorporation of the peptide NYLTHRQ and PPy. Further, the scaffold was shown to form stable complexes and interactions with two cell receptors commonly found in cardiac tissue.
Due to the promising results obtained from simulation studies, as a proof of concept, we prepared decellularized beet leaves, cross-linked with the peptide NYLTHRQ and further attached PPy and then examined their interactions with co-cultures cardiomyocytes and vascular smooth muscle cells to examine if such composites could potentially be utilized as potential tissue engineering materials.

Cell viability and interactions
To examine the viability and interactions of the co-cultures of CMCs and SMCs with the decellularized leaf-peptide-PPy (DC-pep-PPy) scaffolds, we performed cellular assays ( Figure 12). First, the MTT assay was performed. MTT assay is based on formation of purple formazan crystals by reduction of yellow tetrazolium salt catalyzed by dehydrogenase enzymes secreted from the mitochondria of metabolically active, and therefore viable cells (van Meerloo et al., 2011). The results obtained are shown in Figure 12a. As shown, cells continued to proliferate over the period of 11 days. In general, 90%, 93% and 94% proliferation was observed after 4, 7 and 11 days respectively indicative of cell viability in the presence of the scaffolds. A longer period of time with continued proliferation indicates metabolic activity which is essential for the functioning of cardiac cells. In a previous study, the proliferation of H9c2 rat cardiac cells on polycaprolactone-carbon nanotube scaffolds was assessed using an MTT assay after 4 days of cell seeding (Ho et al., 2017) to ensure metabolic activity. In another study, researchers found that CMCs from rats were able to maintain physiological levels of metabolic activity after one week of being cultured on a 3 D scaffold (Bursac et al., 1999). CMCs perform high levels of glycolysis during the early stages of proliferation in order to produce sufficient energy for the rapidly dividing cells (Lopaschuk & Jaswal, 2010). Thus, it is essential to utilize a longer time period for the MTT assay to ensure that the cell viability assay measures metabolic activity which is critical to CMC proliferation. To examine the interactions of the cells with the scaffolds, we also performed SEM imaging.
To examine the influence of the scaffold on myofibril development, we performed alpha-actinin antibody staining assay. a-Actinin is a vital regulator of cell adhesion, proliferation and maturation (Choi et al., 2008). Furthermore, it has been shown that weak association of actinin with integrin is necessary for early cell-adhesions (Manisastry et al., 2009). Thus, important information regarding cell-scaffold binding interactions and normal cardiomyocyte formation can be obtained using this assay. In a study assessing cell proliferation and function with a bioactive chitosan nanofiber scaffold, the results indicated that a co-culture of CMCs and fibroblasts resulted in retained morphology and higher expression of proteins such as alpha-sarcomeric actin when compared to cultures of CMCs with the scaffolds. Thus, the co-cultures preserved the proper morphology, spatial organization, and contractility when compared to a monoculture of CMCs (Hussain et al., 2013). Results obtained after incubation with DC-pep-PPy scaffold after one week of growth ( Figure 12b) demonstrate the alpha-actinin cytoskeletal arrangement of the CMCs and SMCs. The results show that the cells formed striations of alpha-actinin filaments indicating differentiation of cardiomyocytes. These results further confirm that the scaffolds supported the growth and maturation of the cells.
Troponin T immunofluorescence assay (Figure 12c) was performed to determine if the cells were continuing to display typical cardiomyocyte markers such as Tropnin cnT indicative of continuous maturation of cells. Thus, cell-seeded scaffolds were stained with m-IgGj BP-CFL 488 antibody and the assay was performed. Results obtained showed that the cells were found to show unique striations typically seen in myocytes. Furthermore, they were in organized, parallel alignments indicated by the arrow. These data indicate that normal cells processes are occurring, and that the scaffolds are interacting with co-cultured cells, allowing for typical sarcomere formation. Additionally, the presence of troponin T indicates the contractile phenotype of CMCs, and is essential to proper cardiac function

Conclusions
In this work, a novel decellularized leaf-peptide-polypyrrole scaffold was both modeled and fabricated for potential applications in cardiac tissue engineering. The molecular dynamics simulations indicated formation of the scaffold consisting of cellulose, pectin, NYLTHRQ, and PPy. Mechanical properties obtained from molecular simulations studies conducted showed that the scaffolds displayed Young's Modulus comparable to scaffolds that may be useful in tissue engineering. Molecular simulation studies further indicated that the scaffold interacted with and successfully assembled in the presence of integrin receptor as well as NPR-C, but appeared to have a higher affinity and stability with NPR-C receptor. These results indicated that the scaffolds may adhere to typical cells commonly found in cardiac tissue. Experimentally, as a proof of concept, decellularized beet leaves were prepared and cross-linked with the peptide NYLTHRQ and then attached to polypyrrole. It was found that the beet leaves were successfully decellularized, leaving minimal amounts of native protein and DNA. Formation of the scaffold and incorporation of peptide NYLTHRQ and polypyrrole was confirmed using FTIR and SEM analysis. Co-cultured cardiomyocytes and vascular smooth muscle cells adhered to, and proliferated in the matrices. MTT assay demonstrated that the scaffolds promoted cell proliferation. Troponin T and actinin assays showed the presence of cardiac specific markers indicative that the cell-scaffold matrices encouraged the proliferation, and growth of cardiac specific cells. Such plant based hybrid scaffolds may be desirable as a promising biomaterial for the potential regeneration of cardiac tissue or may be modified for use in other types of tissue engineering such as neuronal tissue engineering. Future studies may involve transplantation of the biomaterials to examine the effects on cardiac functionality in vivo. Furthermore, such plant based biohybrids may hold therapeutic potential as a highly economic, renewable source for preparation of cardiac patches for tissue regeneration.